Abstract:  The income effect of Chinese households’ social donation behavior is re-examined. Using data from Chinese Family Panel Studies in 2014, 2016 and 2018, truncated regression and Heckman two-stage model are employed to test and solve possible truncation and sample selection bias problems in OLS (ordinary least square) regression. It is found that the proportion of households’ participation in social donations in China is 19.1%, and the average annual donation in these households is 449.61 yuan. The income elasticity of social donations is 0.29, and the difference from the result of the OLS (0.21) is mainly caused by left truncation. Due to the lower participation rate of middle-income households, social donations show a clear “U”-shaped income relationship. The three modes of households’ participation in tertiary distribution, i.e., gifts, financial support and social donations are compared and the reasons for the differences in the income effects of the three modes are analyzed. Based on the above discussion, it is proposed that channels for households’ participation in tertiary distribution should be further broadened.

Key words: tertiary distribution, social donation, income effect, common prosperity, truncated regression